Study And Application On Spatial Database Index

As the information load of network device is getting increasingly huge, it is ratherdifficult to manage in an effective and convenient manner by utilizing the traditional dataand methods. Therefore, with GIS in the network management system, huge spatialgeographic information can be better dealt with in an efficiently precise and logical way.Given that there is little study on such area and GIS may well be the trend for the futuredevelopment of network management system, such study can play a practically importantrole in network management.This paper firstly introduces a clustering algorithm of CHAMELEON based oncompressed data to handle the exponentially growing spatial data and complicated structure.This algorithm can deal with high dimensional data in an automatic andefficient manner. First, combine the given data aggregation and the nearest data, andconstruct V diagram by iteration until the data size dramatically decreases. Then carry onCHAMELEON clustering to the compressed data. Meanwhile, throughsimulation experiments, verify the size of compressed data is far smaller. Thus, the wholeclustering process gets highly efficient.Secondly, the author puts forward a new algorithm with double R-tree (DR-tree) tosolve the issues caused by tremendous data size, deepening depth of R-tree, increasinglylarge occupied space and low query efficiency. He also gives out the data structure, thealgorithm of insert, index and delete. Based on simulation experiments, DR-tree will supplyhigher efficiency at the expense of slightly more occupied space. Moreover, the more theindex objectives, the better retrieval performance can be delivered.Finally, the author describes network management system and its necessity. Byillustrating the basis of choosing the development tool, he selects MapInfo to manage andapply to the network management system.